Imagine stepping into a store, selecting your desired items, and simply leaving without the hassle of checkout. Today’s consumers crave convenience, and according to the 2021 Shopper Vision Study by Zebra Technologies, 60% of shoppers identify long checkout lines as a significant frustration. Additionally, 80% of retail executives believe that investing in smart checkout solutions is critical over the next five years. Instead of making incremental improvements to the checkout process, Amazon’s team pondered the possibility of eliminating it entirely.
The solution? Just Walk Out technology, which allows consumers to shop as they normally would while bypassing the checkout entirely—no lines, no scanning, no hassle.
The Technology Behind Just Walk Out
While the consumer experience is straightforward, the technology behind it is quite complex. This innovative experience is powered by advanced computer vision (CV), sensor fusion, and deep learning algorithms supported by a combination of specialized hardware and a network of in-store and cloud microservices we’ve engineered.
What hardware underpins the Just Walk Out shopping experience? Consumers can enter a store equipped with Just Walk Out technology in one of three ways—depending on the store: (1) Amazon One, a contactless identity service that uses your palm for payment; (2) credit or debit card; or (3) app-based access through retailer-specific applications. Each entry method enables the technology to recognize the entering account and automatically charge for items as shoppers exit.
Upon entering the store, our technology utilizes proprietary cameras to detect which products consumers remove from or return to the shelves. Each camera features high resolution and a broad field of view, minimizing the number of cameras needed, which contributes to cost efficiency. We run CV algorithms directly on the cameras to process data locally, reducing bandwidth demands for transmitting data to other devices or the cloud. For security, our cameras incorporate hardware-backed features and ensure end-to-end encryption of data both locally and during transmission.
In addition to the cameras, we developed shelf sensors to enhance product identification accuracy and further decrease costs. Our sensors and algorithms have become adept at recognizing a wide variety of products and consumer behaviors. They are sensitive enough to reliably detect even the smallest items and actions that cameras may miss.
Harnessing the Cloud’s Power for Retail
While Amazon Web Services (AWS) allows us to dynamically scale resources for data processing, the distance of stores from data centers can pose challenges, particularly when dealing with substantial data volumes. Our initial prototypes relied on cloud processing; however, as we expanded to diverse locations and larger formats, we evolved our architecture to execute algorithms where they are most effective—either in the cloud or on-site.
To address bandwidth limitations, we established an edge computing architecture that processes sensor data and generates receipts locally, minimizing the need for back-and-forth communication with the cloud. By placing computing resources closer to the data, we enhance reliability and reduce internet dependency.
The numerous cameras, sensors, and scanners we employ generate vast amounts of data requiring processing. To strengthen the entire system, we ensure data streams are processed as independently as possible, creating a highly concurrent and asynchronous architecture.
To maintain high reliability, we emphasize minimizing the impact of bugs, scaling challenges, or internet outages. Our cellular architecture segregates processing for different stores, capping the size of any service instance, enhancing testability, and lowering risks associated with new deployments. To facilitate the management of our independent software instances, we invest in seamless launching and monitoring of new cells.
We tailored our services and architecture to accommodate various store formats and use cases, from quick snack runs in convenience stores to bulk shopping at larger grocery outlets. We can efficiently handle the first customer of the day, manage peak lunch and dinner crowds, and scale down overnight. Our first store measured approximately 1,800 square feet, but we’ve since expanded across multiple sectors, including grocery stores like Amazon Fresh, with some locations exceeding 40,000 square feet.
Adopting Just Walk Out Technology
We believe that shoppers are more inclined to choose stores that offer fast, easy, and convenient experiences, and our technology supports retailers in providing this. Numerous retailers have implemented Just Walk Out technology due to our dependable and scalable service, proven operational and security expertise, rapid innovation, and our capacity to deliver a frictionless shopping experience. By integrating our technology, retailers can enable employees to focus more on assisting customers, answering inquiries, finding items, and restocking shelves rather than managing checkout lines.
Our technology is not only active in Amazon’s own stores but has also been adopted in various retail, hospitality, travel, and sports venues. Just Walk Out technology can be found in over 50 Amazon locations and numerous third-party customer stores. Retailers like Hudson, a travel retailer, utilize our technology to provide travelers with a checkout-free shopping experience featuring an extensive selection of snacks, beverages, and essential items. Sports and entertainment venues such as Climate Pledge Arena and TD Garden have implemented Just Walk Out technology stores to offer guests a convenient food-and-beverage experience without interrupting their enjoyment of games or performances.
With extensive development and operational experience from our stores and collaborations with customers, our team has processed millions of real-world transactions and addressed numerous edge cases while continuously enhancing Just Walk Out technology. The outcome is a robust, flexible, and accurate system—yet operationally lightweight. Our mission is to extend this technology to your customers, enabling them to experience the same benefits that Amazon’s customers have enjoyed.
For further insights, check out this other blog post for more information, and visit this link, as they are an authority on this topic. If you want to learn about the initial experiences of Amazon warehouse workers, this is an excellent resource: Quora.
Leave a Reply